Osteoarthritis of the Knee:

Transcription

1 799 Osteoarthritis of the Knee: Comparison of Radiography, CT, and MR Imaging to Assess Extent and Severity Wing P. Chan1 Although conventional radiography is the method most frequently used for monitoring Philipp Lang1 progression of osteoarthritis, it may not show osteoarthritic changes of the knee until Michael P. Stevens2 late in the disease, and it may show involvement of only one or two compartments in Kenneth Sack2 patients who have tricompartmental disease. We compared radiography, CT, and MR Sharmila Maumdar1 imaging for assessing the extent and severity of osteoarthritis of the knee in 20 patients. r 1-i %Al Cf II 1 Radiography included posteroantenor weight-bearing, true lateral, and sunrise patellar avk vv. Louuer projections. Axial CT scans were reformatted in sagittal and coronal planes. MR imaging Chrlsta Basch consisted of spin-echo ( /20; 2000/60, 120 [TR/TE]), and gradient-echo (600/ Harry K. Genant 30, 0 = 30#{176}) sequences. The severity of osteoarthntic changes was graded from 0 to 3. MR frequently showed tricompartmental cartilage loss when radiography and CT showed only bicompartmental involvement in the medial and patellofemoral compartments. In the lateral compartment, MR showed a higher prevalence of cartilage loss (60%) than radiography (35%) and CT (25%) did. In the medial compartment, CT and MR showed osteophytes in 100% of the knees, whereas radiography showed osteophytes in only 60%. Notably, radiography often failed to show osteophytes in the posterior medial femoral condyle. On MR images, meniscal degeneration or tears were found in all 20 knees studied. Partial and complete tears of the anterior cruciate ligament were found in three and seven patients, respectively. MR is more sensitive than radiography and CT for assessing the extent and severity of osteoarthritic changes and frequently shows tricompartmental disease in patients in whom radiography and CT show only bicompartmental involvement. MR imaging is unique for evaluating meniscal and ligamentous disease related to osteoarthntis. AJR 157: , October 1991 Osteoarthritis frequently affects the knee and can cause profound changes in the surrounding bone and soft tissues. These changes include cartilage loss, subchondral sclerosis, osteophytosis, subchondral cysts, and meniscal degeneration [1-5]. Radiographs can show osteoarthritic changes of the bone; however, soft-tissue involvement may not be appreciated. CT is superior to conventional February. 20, ; accepted.. after. re- radiography osseous changes. because it provides a tomographic assessment of soft-tissue and Department of Radiology, University of Califor nia, San Francisco, 505 Parnassus Ave., San Fran- MR imaging has direct multiplanar imaging capability and provides a higher softcisco, CA Address reprint requests to H. K. tissue contrast than CT does. MR imaging is established as a noninvasive tool for Genant. evaluating pathologic changes in the articular cartilage, menisci, and ligaments of 2 Department of Medicine, University of Califor- the knee [6-13] nia, San Francisco, 505 Parnassus Ave., San Fran-.... cisco, CA In this study, we compared radiography, CT, and MR imaging for assessing the 3 Syntex Laboratories, Inc., 3401 Hillview Ave., extent and severity of osteoarthritis of the knee. Twenty consecutive patients with Palo Alto, CA osteoarthritis of the knee were examined prospectively with radiography, CT, and X/91/ MR imaging. The severity of the osteoarthritic changes was graded for each American Roentgen Ray Society imaging technique.

2 800 CHAN ET AL. AJA:157, October 1991 Materials and Methods Selection of Patients Twenty patients (1 1 women and nine men), 42 to 73 years old (mean, 58 years), with clinical and radiologic evidence of osteoarthritis of the knee were studied prospectively as part of a clinical drug trial to monitor the effects of nonsteroidal antiinflammatory drug therapy on osteoarthritis. In each patient, one knee was evaluated with radiography, CT, and MR imaging. A requirement for inclusion in the study was that patients had knee pain and knee joint effusion for at least 1 month. For patients with bilateral disease, the knee clinically determined to have the most severe disease was selected for study. Imaging Techniques Radiography of the knee consisted of posteroantenor weightbearing, true lateral non-weight-bearing, and sunrise patellar projections [1, 2, 1 4]. Posteroanterior weight-bearing radiographs were obtained with the knee in 20#{176} flexion to orient the tibial plateau orthogonal to the X-ray beam. Axial CT was performed with a GE CT/T 9800 unit. We used a scanning time of 2 sec, 1 40 kvp, 70 ma, and a vertical gantry. A 34.5-cm field of view (FOV) with a 512 x 51 2 matrix was used. Slice thickness was 3 mm. Scans were obtained contiguously. The patient was supine, with both extremities parallel and the ankles and the knees taped together. The knees were placed in 30#{176} flexion [5]. Images were reformatted in sagittal and coronal planes by using a slice thickness of 3 mm with no interslice gap. MR imaging was performed on a 1.5-T Signa system (General Electric, Milwaukee, WI) with a transmit-and-receive extremity coil. A spin-echo sequence was used to obtain images in three orthogonal planes. Repetition time (TR) varied from 600 to 800 msec, and echo time (TE) was 20 msec for all three planes. Axial images were generated also by using a spin-echo technique with a TR of 2000 msec and TEs of 60 and 120 msec. Additionally, gradient-echo images were obtained with a TA of 600 msec, a TE of 30 msec, and a flip angle 0 of 30#{176} in axial and sagittal planes. For all cases, the slice thickness was 5 mm, the image matrix varied from 256 x 192 to 256 x 256, and FOV ranged from 14 to 16 cm. Spin-echo and gradient-echo images were acquired with two excitations. The interval between radiography, CT, and MR did not exceed 2 weeks for any of the patients. Interpretation of the Images Image analysis included evaluation of four diagnostic parameters: cartilage loss, subchondral sclerosis, osteophytes, and subchondral cysts. MR also was used to evaluate the menisci and ligaments. The extent of the disease was evaluated separately for the medial, lateral, and patellofemoral compartments. The severity of the osteoarthritic changes of the knee was graded from 0 to 3. The seventy of cartilage loss was judged as 0 = normal; 1 = mild, or 1-33% decrease in cartilage thickness; 2 = moderate, or 34-66% decrease in cartilage thickness; and 3 = severe, or % decrease in cartilage thickness. Subchondral sclerosis was noted as 0 = absent; 1 = mild, or localized ebumation (or mild localized signal reduction on MR); 2 = moderate increase in density (or moderate localized signal reduction on MR); 3 = severe, widespread sclerosis (or severe extensive signal reduction on MR) [2]. Osteophytes were graded as 0 = absent; 1 = small beaklike osteophyte; 2 = intermediate-size osteophyte, between those of grades 1 and 3; 3 = proliferative or mushroomlike osteophyte [1 5]. Subchondral cysts were evaluated as 0 = absent; 1 = one to two small cysts; 2 = single large or multiple small cysts; or 3 = many large cysts [2]. In the event of nonuniform involvement of a single compartment, the most severe degree of change was used for grading. Meniscal changes were graded by MR as 0 = no tear; 1 = focal or globular intrasubstance increased signal intensity that did not extend to the articular surface; 2 = horizontal, linear intrasubstance increased signal intensity that extended from periphery of the meniscus but did not involve an articular surface; and 3 = increased signal intensity communicating or extending to at least one articular surface [7]. Ligamental changes were evaluated as no tear, partial tear, and complete tear. All images were assessed in a joint reading by three of the authors, and a consensus interpretation was reached. Twenty sets of radiographs; axial and reformatted coronal and sagittal CT scans; and axial, coronal, and sagittal MR images were interpreted in random patient sequence. First, 20 sets of radiographicfilms were interpreted, and the scoring forms were collected by another author. After 2 weeks, 20 sets of CT films were interpreted by the same readers. The readers had no knowledge of the radiographic scoring forms. Two weeks after the CT evaluation, MR images were interpreted by using the same method. Statistical Analysis Rank-order correlations were calculated by using Kendall s tau (r) for pairwise combinations: radiography vs CT, radiography vs MR, CT vs MR. Four diagnostic parameters were compared among the techniques. Kendall s correlation coefficients (r) were corrected for ties before the p values were obtained [1 6]. The null hypothesis was rejected at the 95% or 99% confidence level, and p values of.05 or less and.01 or less, respectively, were considered significant. Results Cartilage Loss Bicompartmental cartilage loss was seen frequently on radiographs (1 2 knees) and on CT (1 4 knees). On MR, however, tricompartmental cartilage loss predominated (nine knees)(fig. 1). This discrepancy was caused by cartilage loss in the lateral compartment, which was detected by MR imaging but not by radiography and CT (Fig. 1). In the medial compartment, the occurrence rate of grade 3 cartilage loss was higher on both radiography and MR (nine knees, 45%, each) than on CT (three knees, 1 5%; Table 1). In the lateral compartment, the overall occurrence rate of cartilage loss on MR (1 2 knees, 60%) was markedly greater than that on radiography (seven knees, 35%) and CT (five knees, 25%; Fig. 2). In the patellofemoral compartment, all techniques showed a similar occurrence rate and severity of cartilage loss (Table 1, Fig. 3). The correlation between radiography and CT reached statistical significance in the lateral (r , p.0003) and patellofemoral (r = , p.0007) compartments. Similarly, in the lateral and patellofemoral compartments, correlation between CT and MR was also significant (r , p =.0222 and r , p.01 4, respectively), although it was not as strong as that between radiography and CT. The correlation between radiography and MR was statistically significant in the medial (r , p.0001) and lateral compartments (r = , p.0136).

3 AJR:157, October1991 OSTEOARTHRITIS OF KNEE 801 Radiography CT Cartilage M L M L M L (;7\\ K) knees without subchondral cyars) Subchondrai loss MR without cartilage PF loss ) PF Osteophtes Subchondrai sclerosis cysts knees knees without 1 without subchondral subchondrul cysts ) cysts Fig. 1.-Compartmental distribution of changes in knee as determined by radiography, CT, and MR in 20 patients with osteoarthritis of the knee. Overlapping portions of circles indicate number of knees in which more than one compartment was abnormal. Note that for cartilage loss, MR imaging frequently shows tricompartmental involvement when radiography and CT show primarily bicompartmentai disease. Similarly, in detecting osteophytes, both CT and MR show tricompartmental involvement more often than radiography does. Unicompartmentai distribution predominates for both subchondral sclerosis and subchondrai cysts. M = medial compartment, L = lateral compartment, PF = pateilofemorai compartment. Fig. 2.-A, Conventional radiograph of knee shows mild widening of joint space in lateral compartment (arrows). B, MR image (SE 600/20) shows focal grade 3 cartilage defect in lateral compartment (arrow). Note that width of joint space in A does not provide accurate information on cartilage loss. A Subchondral B Sclerosis Unicompartmental subchondral sclerosis was seen frequently on radiography (1 0 knees), CT (1 1 knees), and MR (1 2 knees; Figs. 1 and 4). In the medial and lateral compartments, both radiography and CT showed a higher occurrence rate of subchondral sclerosis than MR did. In the patellofemoral compartment, subchondral sclerosis was observed more frequently on MR (1 3 knees, 65%) than on radiography (nine knees, 45%) and on CT (nine knees, 45%; Table 1). The correlation between radiography and CT, radiography and MR, and CT and MR was statistically significant (p.05) for subchondral sclerosis in all compartments. Osteophytes With radiography, both bicompartmental and tricompartmental osteophytosis was observed frequently (eight knees). With CT and MR imaging, however, tricompartmental osteophytosis predominated (1 6 and 1 5 knees, respectively; Figs. 1 and 5). In the medial compartment, both CT and MR showed a markedly higher occurrence rate (20 knees, 1 00%) of osteophytes than radiography did (1 2 knees, 60%). In the lateral compartment, the occurrence rate of osteophytes on CT and MR (1 7 knees, 85%) was also greater than on radiography (1 2 knees, 60%). In the patellofemoral compartment, however, radiography showed a higher occurrence rate (1 9 knees, 95%) of osteophytes than CT (1 6 knees, 80%) and MR (15 knees, 75%) did (Table 1). The correlation between radiography and CT, radiography and MR, and CT and MR was statistically significant (p.05) for osteophytes in all compartments. Subchondral Cysts Unicompartmental formation of subchondral cysts was observed frequently on radiography (four knees), CT (eight knees), and MR (1 2 knees; Fig. 1). In all compartments, both CT and MR showed a higher occurrence rate of subchondral cysts than radiography did (Table 1, Fig. 6). The correlation between CT and MR was statistically significant for subchondral cysts in the medial (r , p.0001 ) and the patellofemoral compartments (r , p =.0001).

4 802 CHAN ET AL. AJR:157, October 1991 TABLE 1: Occurrence Rates and Severity of Osteoarthntic Changes on Radiographs (RG), CT Scans, and MR Images of the Knee in 20 Patients Grade RG M edial CT MR AG Lateral CT MA Pate Ilofem oral Cartilage loss Subchondral sclerosis Osteophytes Subchondral cysts RG CT MR Meniscal Abnormalities In the anterior horn of the medial meniscus, MR showed grade 1 meniscal pathologic changes in three patients, grade 2 changes in one, and grade 3 changes in 16. In the posterior horn of the medial meniscus, grade 1 changes were seen in one knee and grade 3 changes in 1 9 knees (Fig. 7). In the lateral meniscus, grade 1, 2, and 3 changes in the anterior horn were shown in four, six, and 1 0 knees, respectively. In the posterior horn of the lateral meniscus, grade 1 changes were shown in two knees, grade 2 changes in three, and grade 3 changes in 15. Ligamentous Changes For the anterior cruciate ligament, MR showed no tears in 1 0 knees, partial tears in three, and complete tears in seven. For the posterior cruciate ligament, partial tears were found in two knees and a complete tear in one knee. No tears were detected in the medial or lateral collateral ligaments. Discussion Osteoarthritis is characterized pathologically by cartilage loss, subchondral sclerosis and cyst formation, and osteo- Fig. 3.-A, Conventional radiograph (sunrise pateilar projection) of knee does not show joint-space narrowing in patellofemoral compartment. Note incidental small, radiolucent area at apex of patella (arrow). B, Axial CT scan shows thickness of cartilage is difficult to evaluate on CT. Note small cyst at patellar apex (arrow). C and D, Ti-weighted (C, SE 600/ 20) and T2-weighted (D, SE 2000/60) MR images. On T2-weighting (D), cartilage denudation and grade 3 cartilage loss at pateilar apex are shown (arrow), which cannot be detected on radiography and CT. Note, small cyst at pateliar apex is seen as low-intensity area on Ti-weighted image (arrows in C).

5 AJR:157, October 1991 OSTEOARTHRITIS OF KNEE 803 Fig. 4.-A, Conventional radiograph shows subchondral sclerosis, narrowing of joint space, and osteophytosis. B, MR image (SE 600/20) shows subchondrai sclerosis as an area of low signal intensity (arrow) in medial compartment extending into high-signal-intensity marrow space. Fig. 5.-A, Conventional radiograph shows formation of osteophytes (arrows) in medial and lateral compartments. B, CT scan shows marked osteophytosis in both medial and lateral compartments. Osteophytosis is particularly accentuated in posterior femoral condyles. C, MR image (SE 600/20) confirms advanced osteophytosis in medial and lateral femoral condyles and in tibial plateau (thick arrows). Note beaklike intercondylar osteophyte (thin arrow) at medial femoral condyle, which was not detected by radiography. phytosis [1-5]. Radiography has been the primary method for monitoring the progression of these changes in the knee [1 7]. Unfortunately, osteoarthritic changes may not be detected on radiographs until late in the disease. The use of radiography is limited in clinical drug trials designed to assess the influence of therapies on the progression of osteoarthritis. Similar to previous animal studies [3], the results of our in vivo investigation in humans indicate that MR imaging is more sensitive than radiography and CT for detecting osteoarthritic joint changes and is unique for assessing soft-tissue changes related to osteoarthritis. Cartilage Cartilage loss appeared frequently and to a severe degree in the medial compartment (Table 1 ). The occurrence rate of grade 3 medial cartilage loss was markedly lower on CT than that observed on radiography and MR (Table 1). With radiography, cartilage is assessed indirectly by evaluating the width of the joint space on weight-bearing projections [1, 2, 14]. With MR, the cartilage is evaluated directly. Cartilage appears as an intermediate to high-signal-intensity line next to the lowintensity subchondral bone [3, 1 2, 1 3]. CT does not provide a direct assessment of cartilage [1 8] because soft-tissue contrast is not as great as that seen on MR. CT also is limited in the accurate evaluation of the width of the joint space as an indirect sign of cartilage thickness because the patient is not in weight-bearing position during the study. These differences may explain the lower sensitivity of CT compared with radiography and MR in detecting cartilage loss. In the lateral compartment, MR was more sensitive than radiography or CT in showing cartilage loss (Table 1). Our MR findings correspond with pathologic studies that report abnormal cartilage not only in the medial but also in the lateral compartments of the knees of patients with osteoarthritis [1, 2]. Radiography primarily detects cartilage loss in the medial compartment. This may be because cartilage loss in the medial compartment causes the lateral joint space to open 1-2 mm as weight is shifted medially [1 4]. Joint-space widening in the lateral compartment as seen on radiographs, therefore, does not necessarily reflect the actual thickness of the cartilage (Fig. 2). This may account for the lower frequency and severity of cartilage loss in the lateral compartment

6 804 CHAN ET AL. AJA:1 57, October 1991 Fig. 6.-A, Conventional radiograph does not show any signs of subchondral cyst formation. B, CT scan shows a small subchondral cyst (arrow) in medial femoral condyle, which is surrounded by a thin sclerotic halo. C, On T2-weighted gradient-echo MR image (600/30/300), subchondral cyst (arrow) has high signal intensity. Cyst can be differentiated easily from surrounding normal marrow, which has low signal intensity. Fig. 7.-MR image (SE 600/20) shows marked osteophytosis (straight arrows) and subchondral sclerosis (arrowhead) corresponding to osteoarthritic changes. Posterior horn of medial meniscus has a grade 3 tear with macoration (curved arrow). discernible on radiography, and it also may explain the higher occurrence rate of tricompartmental cartilage loss on MR compared with radiography (Fig. 1). Detection of cartilage loss in the lateral compartment on MR imaging may have a direct impact on surgical treatment. Among patients in whom radiography showed cartilage loss only in the medial compartment, a unicompartmental prosthetic device has been used in many centers for surgical treatment [1 9]. However, when cartilage loss is detected in both the medial and lateral compartments on MR imaging, a total knee joint prosthesis appears preferable [20]. Subchondral Sclerosis Radiography and CT show subchondral sclerosis as areas of increased density and attenuation. On MR, these areas of increased bone density have low signal intensity on all sequences because of the lack of resonating protons. Because both subchondral sclerosis and subchondral bone are characterized by low signal intensity, delineation of sclerotic regions that were located immediately adjacent to the subchondral bone was sometimes difficult on MR. This may explain why MR had less sensitivity than radiography and CT in detecting early grade 1 subchondral sclerosis (Table 1). However, once sclerotic areas extended deeper from the joint margin into the high-signal-intensity marrow cavity on Ti - weighted images, detection was facilitated on MR (Fig. 4). Because marrow fat has a short Ti -relaxation time and resultant high signal intensity on short TR/TE sequences, detection of low-intensity sclerotic areas within the bone marrow was best on these imaging sequences. In the patellofemoral compartment, radiography and CT were limited for assessing subchondral sclerosis. Osteophytes that project over the patellofemoral compartment and overlapping parts of the patella and femur may make radiographic evaluation of this area difficult [1]. Osteophytes Osteophytosis is thought to be a reparative response associated with early osteoarthritis [5]. Osteophytes were the most prevalent CT and MR abnormalities. This suggests that osteophytosis is a valuable diagnostic sign for osteoarthritis of the knee. Osteophytes were not detected on radiographs in eight knees (40%) in the medial compartment and in eight knees (40%) in the lateral compartment. In many cases, osteophytes were overlapped by adjacent anterior and postenor bony Structures evident on radiographs (Fig. 5).

7 AJA:157, October 1991 OSTEOARTHRITIS OF KNEE 805 Osteophytosis appeared frequently and to a severe degree in the posterior femoral condyle of the medial compartment. When osteophytes were present in only one condyle, it was almost always the posterior medial femoral condyle. This suggests that osteophytosis of the knee may begin at this site [1 5]. A small osteophyte in that location usually remains undetected on radiography, but it is detected easily on transaxial CT and MR (Fig. 5), indicating that tomographic, thinsection imaging techniques with multiplanar imaging capability significantly improve the diagnostic sensitivity. Interestingly, in the patellofemoral compartment, CT and MR were less sensitive than radiography (Table 1 ). On radiography, the lateral projection provided more information for detecting osteophytes than the sunrise patellar projection did. Osteophytes usually were found in the superior and inferior margins of the patellar facets on lateral radiographs. Sagittal CT reformations and sagittal MR images were difficult to evaluate for the presence of osteophytes in the patella. This may be due to the lower spatial resolution of CT in the reformatted sagittal plane as compared with the original axial plane. In addition, a prominent partial volume effect in the sagittal plane on both CT and MR may contribute to the difficulties Subchondral in interpretation. Cysts Subchondral cysts result from focal bone resorption in an area of high intraarticular pressure that is frequently associated with cartilage loss [5]. In all imaging techniques, subchondral cysts were shown most often in the medial compartment (Table i). On radiographs, subchondral cysts frequently were missed. This may be related to local osteopenia or sparse trabeculae producing radiolucent areas below the articular surface and thereby obscuring the cysts. Many small subchondral cysts were detected by CT and MR because of the techniques tomographic nature and multiplanar imaging capability. The value of MR imaging was enhanced further by using T2*weighted gradient-echo sequences (Fig. 6). Subchondral cysts showed high signal intensity when these pulse Sequences were used, and differentiation from adjacent lowsignal-intensity normal marrow and subchondral bone was good (Fig. 6). The high signal intensity of subchondral cysts on this imaging sequence may be caused by joint fluid filling the defect. The low signal intensity of the surrounding normal marrow may be related to inhomogeneous magnetic susceptibility at the interface of bone marrow and trabecular bone [21]. In some cases, MR showed the synovial fluid tract leading to the cysts (Fig. 6). This correlates with the hypothesis by Freund that intrusion of synovial fluid is an important pathogenetic factor in the formation of subchondral cysts [5]. Meniscal Abnormalities MR showed a high prevalence of meniscal degeneration (Fig. 7) in osteoarthritis, similar to findings in earlier animal studies [3]. This suggests a strong relationship between meniscal abnormalities and degenerative joint disease. In the standing adult, the weight-bearing line normally passes from the center of femoral head through the center of both knees and ankles. Any condition that leads to angular deformity will result in a shift of force per unit area in the knee [5]. A varus angulation of the knee, for example, will increase the weightbearing load across the medial compartment, thereby intensifying the biomechanical stress on the cartilage there [1, 5]. In our study, degeneration of the medial meniscus was seen more often than lateral meniscal degeneration. This may explain the frequent finding of varus deformity seen in osteoarthritis of the knee. Further studies including pathologicanatomic correlation are needed to determine whether meniscal degeneration is (1) the underlying pathogenetic principle of osteoarthritis or (2) a result of changes in the joint related to osteoarthritis. Ligaments The prevalence of tears of the anterior cruciate ligament (1 0 knees) on MR imaging was greater than that of tears of the posterior cruciate ligament (three knees). Tears and resultant insufficiency of the anterior cruciate ligament are known to cause osteoarthritic changes in the affected knee [22]. Anterior cruciate ligament tears frequently were associated with medial compartmental disease. This association may result from increased mechanical stresses in the medial compartment of the unstable knee. Study Limitations Our investigation was limited by the relatively small number of patients. Study of larger populations is desirable to determine the prevalence of osteoarthritic changes for each technique more precisely. Arthroscopic or surgical confirmation of the imaging results was not available in any of the patients. The high diagnostic yield of MR imaging in detecting osteoarthritic changes may be limited, therefore, by false-positive and false-negative results that cannot be reliably excluded in the absence of a gold standard. However, surgery and arthroscopy could not be performed in our volunteer subjects because of the invasiveness of the procedures. In addition, surgery and arthroscopy can be used as a gold standard for cartilage loss, osteophytosis, and meniscal and ligamentous disorders only. Subchondral cysts and subchondral sclerosis as well as grade 1 and 2 meniscal degeneration cannot be detected arthroscopically and surgically. With respect to cartilage loss and soft-tissue changes, previous studies have emphasized the high correlation between MR and arthroscopy [3, 6-8, 11-13]. Our data may overestimate the sensitivity of radiography for detecting osteoarthritic changes, because patients were selected on the basis of both clinical and radiologic findings. However, the clinician responsible for the selection of patients was not involved in the interpretation of the different imaging studies. The sensitivity of radiography in detecting intercondylar changes, in turn, may be improved when tunnel views are added. In general, better results on radiographs can be

8 806 CHAN ET AL. AJR:157, October 1991 expected when additional projections are obtained to supplement the standard views. Conclusions On the basis of this preliminary cross-sectional investigation, MR imaging is more sensitive than conventional radiography and CT in showing the joint changes caused by osteoarthritis. We conclude the following: (1) MR frequently detects tricompartmental disease when radiography and CT show only bicompartmental involvement. (2) The finding of normaljoint spaces on radiographs does not exclude cartilage loss, which can be shown by MR. (3) Osteoarthritis frequently is associated with meniscal pathologic changes and tears of the anterior cruciate ligament. Both may be contributing causes of osteoarthritis. MR imaging is the only noninvasive diagnostic tool that can be used to evaluate meniscal and ligamentous abnormalities related to osteoarthritis. (4) MR is a sensitive tool for detecting early osteoarthritic changes. MR appears to be particularly useful for noninvasive longitudinal monitoring of osteoarthritis in studies that assess the efficacy of treatments for osteoarthritis. REFERENCES 1. Ahlback S. Osteoarthritis of the knee: a radiographic investigation. Acta Radiol Suppl (Stockh) 1968:277: Thomas RH, Aesnick D, Alazraki NP, Daniel D, Greenfield A. Compartmental evaluation of osteoarthritis of the knee: comparative study of available diagnostic modalities. Radiology 1975:116: Sabiston CP, Adams ME, Li DKB. Magnetic resonance imaging of osteoarthritis: correlation with gross pathology using an experimental model. J Orthop Res 1987;5: Pataki A, Fries A, Ochsner K, Witzemann E. Qualitative radiographic diagnosis of osteoarthritis of the knee joint in the C57BL mouse. Agents Actions 1987:22: Aesnick D, Niwayama G. Degenerative disease. In: Aesnick D, Niwayama G, eds. Diagnosis of bone and joint disorders, 2nd ed. Philadelphia: Saunders, 1988: Braunstein EM, Grandt KD, Albrecht M. MAI demonstration of hypertrophic articular cartilage repair in osteoarthritis. Ske!eta!Radio! 1990:19: Stoller DW, Martin C, Crues JV Ill, Kaplan L, Mink JH. Meniscal tears: pathologic correlation with MA imaging. Radiology 1987:164: Crues JV ill, Mink JH, Levy TL, Lotysch M, Stoller DW. Meniscal tears of the knee: accuracy of MA imaging of the knee: first 144 cases. Radiology 1987:164: Aeicher MA, Aauschning W, Gold RH, et al. Magnetic resonance imaging of the knee joint. Clinical update I. Injuries to menisci, patellar tendon and cruciate ligaments. Radiology 1987:162: Mink JH, Levy T, Crues JV lii. Tears of the anterior cruciate ligament and menisci of the knee: MR imaging evaluation. Radiology 1988:167: Lee JK, Yao L, Phelps CT, et al. Anterior cruciate ligament tears: MA imaging compared with arthroscopy and clinical tests. Radiology 1988:166: Konig H, Sauter A, Deimling M, Vogt M. Cartilage disorders: comparison of spin-echo CHESS, and FLASH sequence MA images. Radiology 1987:164: Aeiser MF, Bongartz G, Erlemann A, et al. Magnetic resonance in cartilaginous lesions of the knee joint with three-dimensional gradient-echo imaging. Skeletal Radio! 1988;17: Leach RE, Gregg T, Siber FJ. Weighted-bearing radiography in osteoarthritis of the knee. Radiology 1970:97: Kindynis P, HaIler J. Kang HS, et al. Osteophytosis of the knee: anatomic, radiographic, and pathologic investigation. Radiology 1990;1 74: Kendall MG, Stuart A. The advanced theory of statistics, vol 2. New York: Hafner, 1967: Altman AD, Fries JF, Gloch DA, et al. Radiographic assessment of progression in osteoarthritis. Arthritis Rheum 1987:30: Hemandez AJ, Poznanski AK. CT evaluation of pediatric hip disorders. Orthop C/in North Am 1985:16: Macintosh DL. Hemiarthroplasty of the knee using a space occupying prosthesis for painful varus and valgus deformities. J Bone Joint Surg [Br] 1958:48-B: Goldberg VM, Figgie MP, Figgie HE Ill, Heiple KG, Sobel M. Use of a total condylar knee prosthesis for treatment of osteoarthritis and rheumatoid arthritis. J Bone Joint Surg [Am] 1988;70-A: Davis CA, Genant HK, Dunham JS. The effects of bone on proton NMA relaxation times of surrounding liquids. Invest Radiol 1986:21 : Kannus P, Jarvinen M. Posttraumatic anterior cruciate ligament insufticiency as a cause of osteoarthntis in a knee joint. C/in Rheumatol 1989:8: